DataPulse Analytics SaaS
Dashboard analityczny w czasie rzeczywistym z interaktywnymi wykresami, raportowaniem rolowym i AI insights dla e-commerce.
- Branża
- E-commerce analytics SaaS
- Czas trwania
- 9 months
- Moja rola
- Fullstack architect
- Client
- DataPulse Analytics
Wyzwanie
An analytics SaaS for e-commerce operators was choking on Postgres aggregations — dashboards took 30+ seconds to render at 1B+ events/month. The product team also wanted AI-powered insights (anomaly detection, forecasts) without rebuilding the stack.
Rozwiązanie
Migrated the analytics warehouse to ClickHouse for sub-second OLAP queries, rewrote the dashboard on Next.js 14 App Router with React Server Components for instant first paint, added tRPC for end-to-end type safety, and wired an OpenAI-backed insights pipeline that flags anomalies and writes plain-language summaries.
Stack technologiczny
Rezultat
Dashboard TTI dropped from 30s to under 400ms, the platform now ingests 1.2B events/month on the same infra, and AI summaries became the #1 retention feature in customer interviews.
Kluczowe metryki
- 1.2B
- events/month
- < 400ms
- dashboard TTI
- AI
- anomaly detection
Dashboard TTI from 30s to under 400ms. AI summaries became the #1 retention feature in customer interviews.
